Replace Sample vs Population

[Mwiinga Ceelo]

Hello, i need help. I have 4 variables i.e.
VRegS VRegP VregPmake and VRegSmake.
Capital S stands for sample while P stands for population

VRegS=vehicle registration for sample

VRegP= vehicle registration for population

VregPmake=...make of a vehicle from the population

VRegSmake=...make of a vehicle from sample
Sample size is 100
Population is 1000

So, I have 100 observations VRegS, 1000 observations of VRegP, and 1000 observations VRegPmake.

And
100 missing values of VRegSmake

So, I want to replace VRegSmake
How?

Here is the catch:
The 100 makes for S can be found in the 1000 makes for P.

I can't use

. replace VRegSmake = VRegPmake if VRegS == VRegP

Because S is smaller than P and the values are not matched one to one, even if I sort, you getting the point?

So, what should I do?

[Mwiinga Ceelo]

hellooo someone help me please

[Clyde Schechter]

you getting the point?

No. In fact, I'm totally confused about what you want to do. I suspect others are as well, which is probably why you got no response. I suggest you post a small representative sample of your data, and then show us a hand-worked solution for that sample so we can see what you want. Please use -datatex- to post your data sample so that those who might want to help you can quickly, easily, and completely faithfully reproduce your example and experiment with it. Run -ssc install dataex- to get the -dataex-command if you don't already have it. Simple instructions for using it are in -help dataex-. I also recommend you read the FAQ for some good suggestions about how to ask questions clearly.

[Mwiinga Ceelo]

All thanks. I have pasted an example of what I am dealing with below.

idS	VRegS	VRegS_make	idP	VRegP	VRegP_make
1	BAB2010	.	1	BAB2010	TOYOTA
2	AAB300	.	2	G0123BT	FORD
3	ACK4500	.	3	AAB300	SCANIA
4	BAC200	.	4	BAA290	MAZDA
5	ADD4205	.	5	ACK4500	BMW
			6	ALM2077	LEXUS
			7	BAC200	MITSUBISHI
			8	ACF500	LOTUS
			9	ADD4205	AUDI
			10	BBA2240	OPEL

Want I want is to replace VRegS_make using the makes for VRegP_make. How? All VRegS is found in VRegP, and VRegP has the VRegP_make (needed to populate VRegS_make).

The challenge is I have 3,054 VRegS observations and 647,521 VRegP observations. Thus, it's not as easy as simply getting rid of the "5" extra observations in the pasted example.

The pasted dataset will thus only be used as a guide for me to work on the larger dataset.

Please help me.

[Mwiinga Ceelo]

Please find attached dataset in stata format. Thanks.

[Jorrit Gosens]

You would be better of keeping separate info in separate datasets in this case, and merging them only in a way that keeps info in each row with the observation they belong to

*move population data to different set
preserve
drop ids vregs vregs_make
ren vregp vregs
save "C:\XYZ\P.dta", replace
restore
*remove population data from sample set
drop vregs_make idp vregp vregp_make
save "C:\XYZ\S.dta", replace
*get rid of missings from sample set
drop if ids ==.
*reintroudce the vreg info properly, thourgh merging them on vreg info
merge 1:1 vregs using C:\XYZ\P.dta